Hospital costs associated with surgical site infections in general and vascular surgery patients

Hospital costs associated with surgical site infections in general and vascular surgery patients Melissa M. Boltz, DO,a,b Christopher S. Hollenbeak, PhD,a,b,c Kathleen G. Julian, MD,d Gail Ortenzi, RN, BSN,a and Peter W. Dillon, MD,b Hershey, PA

Background. Although much has been written about excess cost and duration of stay (DOS) associated with surgical site infections (SSIs) after cardiothoracic surgery, less has been reported after vascular and general surgery. We used data from the National Surgical Quality Improvement Program (NSQIP) to estimate the total cost and DOS associated with SSIs in patients undergoing general and vascular surgery. Methods. Using standard NSQIP practices, data were collected on patients undergoing general and vascular surgery at a single academic center between 2007 and 2009 and were merged with fully loaded operating costs obtained from the hospital accounting database. Logistic regression was used to determine which patient and preoperative variables influenced the occurrence of SSIs. After adjusting for patient characteristics, costs and DOS were fit to linear regression models to determine the effect of SSIs. Results. Of the 2,250 general and vascular surgery patients sampled, SSIs were observed in 186 inpatients. Predisposing factors of SSIs were male sex, insulin-dependent diabetes, steroid use, wound classification, and operative time (P < .05). After adjusting for those characteristics, the total excess cost and DOS attributable to SSIs were $10,497 (P < .0001) and 4.3 days (P < .0001), respectively. Conclusion. SSIs complicating general and vascular surgical procedures share many risk factors with SSIs after cardiothoracic surgery. Although the excess costs and DOS associated with SSIs after general and vascular surgery are somewhat less, they still represent substantial financial and opportunity costs to hospitals and suggest, along with the implications for patient care, a continuing need for cost-effective quality improvement and programs of infection prevention. (Surgery 2011;150:934-42.) From the Division of Outcomes Research and Quality, Department of Surgery,a,b Penn State Milton S. Hershey Medical Center, Department of Public Health Sciences,c and Department of Infectious Disease,d Penn State College of Medicine, Hershey, PA

THE UNITED STATES has an estimated 2.1 million nosocomial infections per year with postoperative surgical site infections (SSIs) being the third most common.1 Despite increased implementation of quality improvement initiatives, SSIs continue to be an ongoing problem. In 1970, the Centers for Disease Control and Prevention’s National Nosocomial Infection Surveillance (NNIS) System was the first to established criteria to classify SSIs as incisional (skin or subcutaneous tissue involvement) or organ-space (any body part other than incised layers involved in the operation). The NNIS also developed a specific risk index using the American Society of Anesthesiologists (ASA) Accepted for publication April 22, 2011. Reprint requests: Christopher S. Hollenbeak, PhD, Department of Surgery, Penn State College of Medicine, 600 Centerview Drive, A210. Hershey, PA 17033-0850. E-mail: chollenbeak@ psu.edu. 0039-6060/$ - see front matter Ó 2011 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2011.04.006

934 SURGERY

physical status classification system, wound classification, and operative duration.2 The NNIS reported that, from January 1992 through June 2004, patients undergoing general surgical procedures including appendectomy, cholecystectomy, colon, gastric, and small-bowel procedures, as well as laparotomy, had a risk of developing an SSI of 2.5% if they had 1 risk factor or 6.2% if they had greater than 2 risk factors.2 Several recent studies report rates of postoperative wound infections between 1.8% and 6.9%, depending on the surgical procedure.3,4 with extensive research conducted on patient and intraoperative risk factors. In addition to increasing morbidity and mortality, SSI is well recognized as one contributing factor to prolonged hospital duration of stay (DOS) and causes substantial financial burden to healthcare systems.5 Policy changes implemented in 2007 by the Center for Medicare and Medicaid Services eliminated reimbursement to hospitals for extra costs associated with injuries or conditions that could have been reasonably prevented,

Surgery Volume 150, Number 5 such as SSIs.6 As a result, hospitals now have a greater financial incentive to decrease postoperative surgical complications.7 Although much has been written about estimated excess cost and DOS associated with SSIs after cardiothoracic surgery, less has been reported on noncardiac procedures. The objective of this study was to use institutional data from the National Surgical Quality Improvement Program (NSQIP) combined with a single-center accounting database to identify risk factors associated with postoperative SSIs and to study the excess total cost and DOS associated with SSIs in patients undergoing general and vascular surgery. METHODS Operations. Of 11,250 adult general and vascular inpatient operations performed between 2007 and 2009 at a single academic medical center, 2,250 were selected for analysis using previously reported NSQIP methodology.8-10 In brief, these cases were selected because they were the first 40 eligible vascular and general surgery operations in an 8-day cycle. Because each cycle began on a different day of the week, these cases captured approximately 20% of the general and vascular surgery volume of the hospital. In all cases, preoperative antibiotics were routinely given within 60 minutes before incision (>90% compliance with the Centers for Medicare and Medicaid Services Surgical Infection Prevention parameters). Shaving was not avoided. The most commonly used preoperative skin antiseptic agent was a commercially available solution composed of Iodine Povacrylex (0.7% available iodine) and Isopropyl Alcohol (74% w/w) (DuraPrep Surgical Solution; 3M Health Care, St. Paul, MN). Patient data. Patient data were collected by a clinical nurse reviewer on 60 preoperative patient characteristics, 18 intraoperative factors, and 22 postoperative occurrences up to 30 days after the procedure. To maintain data reliability, the clinical nurse reviewer completed in-depth training on study definitions, and participated in conference calls, annual meetings, and site visits. Chart review, information from morbidity and mortality conferences, and communication with patients by telephone or letter completed the 30th postoperative day data collection, as described previously.3,8,10,11 SSIs sunder the NSQIP program are defined as superficial, deep, or organ-space, with criteria based on the following definitions established by the Centers for Disease Control and Prevention’s NNIS system. Superficial SSIs involve the skin or subcutaneous tissue of the incision, whereas deep SSIs involve fascia

Boltz et al 935

and muscle layers. Operations involving areas in and around the body organs that become infected are identified as organ-space SSIs. By definition, the infection must have occurred within 30 days postoperatively and only involve tissue that was manipulated or opened during the procedure to be classified as an SSI.12,13 The clinical nurse reviewer assigned the SSI variable based on medical documentation of any of the following: purulent drainage, positive wound culture, or diagnosis of SSI by the surgeon or attending physician. In addition, the superficial SSI variable was assigned if medical documentation indicated redness, swelling, or tenderness of the incision, and the surgeon deliberately opened the wound. Likewise, a deep or organ-space SSI variable was assigned if radiologic studies or re-operation identified an abscess or other evidence of infection at the fascia, muscle, or organ level. Stitch abscesses and infected burn wounds were not classified as SSIs under the NSQIP criteria.12,13 Cost data for the surgical admissions were obtained from the clinical cost accounting database used in our institution (McKesson, San Francisco, CA). These data represented fully loaded operating costs estimated in the database from a ratio of costs-to-charges (CCR) methodology. The CCR approach is a standard, costaccounting technique, whereby costs are estimated as a percentage of hospital charges. Each hospital department estimates a CCR on a yearly basis; the following year, department-level charges are multiplied by the CCR and used as an estimate of costs. The cost for the admission is the sum of all department-level costs. Although the CCR approach is imperfect for estimating hospital costs, it is by far the most prevalent technique. Previous research14 has demonstrated that department-level CCRs are a reasonable approach for estimating costs within a hospital and for estimating costs across hospitals for patients within a common diagnosis related group. Statistical analysis. The statistical analysis was designed to identify risk factors for developing SSIs in this surgical population and to estimate the excess cost and DOS associated with SSIs after controlling for other potentially confounding patient and surgical variables. Patient characteristics were compared using the Student t test for continuous variables and the chi-square test for binary and categorical variables. Logistic regression was used to determine which patient and preoperative NSQIP variables influenced development of SSIs. Total cost and DOS were fit to linear regression models to determine the effect of SSIs after

936 Boltz et al

Surgery November 2011

Table I. Characteristics of infected vs uninfected patients Variable Category General surgery Vascular surgery Sex Female Male Race/Ethnicity White Nonwhite Preoperative DOS, d BMI, kg/mm2 BMI < 18.5 BMI 30+ Nondiabetic Diabetic: insulin-dependent Diabetic: noninsulin-dependent Smoker Radiotherapy Open wound Steroid use ASA class 1 ASA class 2 ASA class 3 ASA class 4 or 5 Wound class Clean Clean-contaminated Contaminated Dirty Operative time, h <4 4–7:59 8–11:59 12+ Emergent operation

No infection, mean/% (N = 2,064)

SSI, mean/% (N = 186)

P value*

0.84 0.16

0.95 0.05

<.0001y <.0001y

0.56 0.44

0.46 0.54

.0140z .0140z

0.95 0.05 0.90 41.03 0.02 0.55 0.82 0.07 0.12 0.21 0.02 0.09 0.05 0.02 0.33 0.58 0.08

0.95 0.05 1.30 41.49 0.04 0.53 0.77 0.12 0.11 0.25 0.06 0.13 0.06 0.01 0.26 0.64 0.09

.6550 .6550 .0530x .8654 .1370 .5630 .1380 .0070z .7350 .1380 <.0001y .1000 .5170 .6110 .0800 .1100 .7340

0.32 0.46 0.12 0.11 3.99 0.67 0.26 0.05 0.03 0.12

0.16 0.54 0.19 0.11 5.29 0.50 0.30 0.16 0.04 0.10

<.0001y .0320x .0060z .9030 <.0001y <.0001y .1820 <.0001y .4570 .3930

*Variables analyzed with a P value > .2 that are not shown in this Table include the following: age > 18 y; BMI, 18.5–30 kg/m2; history of chronic obstructive pulmonary disease; hypertension; acute renal failure; chemotherapy; and previous operation within 30 d. yDenotes statistical significance of P < .0001. zDenotes statistical significance of P < .01. xDenotes statistical significance of P < .05. SSI, Surgical site infection; DOS, duration of stay; BMI, body mass index; ASA, American Society of Anesthesiologists.

adjusting for patient characteristics that may have independently affected costs and DOS. All statistical analyses were performed using commercially available software (Stata Statistical Software: Release 11; StataCorp LP, College Station, TX). Statistical significance was defined as P < .05. RESULTS Demographics. According to the diagnosis related group for admission, the majority of patients underwent operations on the colon and rectum (26.2%) or the foregut (15.9%) and bariatric

surgery (11.5%). Vascular procedures comprised the next largest group of procedures (9.0%). Of the 2,250 surgical inpatients selected for this study, 186(8.2%) developed an SSI. Patient characteristics stratified by occurrence of SSI are depicted in Table I; the groups were similar in demographic characteristics, such as age (P = .57) and race (P = .66). The groups were also alike with regard to preoperative characteristics including body mass index (P = .86), tobacco abuse (P = .14), steroid use (P = .52), and serum albumin concentration (P = .67).

Boltz et al 937

Surgery Volume 150, Number 5

Table II. Logistic regression of factors affecting likelihood of surgical site infection 95% Confidence interval Variable (N = 2,250) Category General surgery Vascular surgery Sex Female Male Ethnicity White Nonwhite Nondiabetic Diabetic: insulin-dependent Diabetic: noninsulin-dependent Nonsmoker Smoker No previous history of steroid use Steroid use ASA class 1 and 2 ASA class 3 and 4 Wound class Clean Clean-contaminated Contaminated or dirty Operative time, h <4 4–8 >8 Nonemergent operation Emergent operation

Odds ratio

Lower

Upper

P value*

Reference 0.20

0.09

0.44

<.0001y

Reference 1.51

1.10

2.07

.0100z

0.57

2.28

.7180

1.18 0.53

3.32 1.45

.0090z .6040

0.97

2.05

.0680

0.69

2.44

.4260

0.90

1.93

.1510

0.88 1.10

2.25 3.02

.1490 .0200x

1.25 2.06

2.58 5.04

.0020z <.0001y

0.38

1.13

.1310

Reference 1.14 Reference 1.98 0.87 Reference 1.41 Reference 1.29 Reference 1.32 Reference 1.41 1.82 Reference 1.79 3.23 Reference 0.66

*Variables analyzed with a P value > .2 that are not shown in this Table include the following: age and whether the patient was admitted from their residence or from another institution. yDenotes statistical significance of P < .0001. zDenotes statistical significance of P < .01. xDenotes statistical significance of P < .05. ASA, American Society of Anesthesiologists.

Yet, some significant differences between the groups were evident, suggesting risk factors for SSI. Of the group with SSIs, 54% were male versus 44% of the control group (P = .014), and 12% of the SSI group had insulin-dependent diabetes versus 6.6% of the control group (P = .007). Several intraoperative variables including clean (P < .001), clean-contaminated (P = .032), and contaminated (P = .006) wound classes, as well as operative times between 8 and 12 hours (P < .001) were also significant between the 2 groups. Risk of infection. The logistic regression results presented in Table II indicate several factors were important in predicting risk of SSI. Male patients were at greater risk of developing SSIs (odds ratio [OR], 1.5; P = .01), as were insulin-dependent diabetics (OR, 1.98; P = .009). Patients with wound

classification of contaminated or dirty (OR, 1.82; P = .02) and operative times between 4 and 8 hours (OR, 1.79; P = .002) or greater than 8 hours (OR, 3.23; P < .0001) were also more likely to develop SSIs. None of the other variables were statistically significant, including factors such as smoking (OR, 1.41; P = .068) and steroid use (OR, 1.29; P = .426). Mortality. Of the 186 patients who developed an SSI, 4 died, which was not statistically significant (P = .847). Interestingly, all 4 patients who died had developed organ-space infections, which was different (P = .019) compared to the 14 patients with deep incisional infections (P = .560) and the other 123 patients with superficial incisional SSIs (P = .076).

938 Boltz et al

Surgery November 2011

Table III. Logistic regression of factors affecting likelihood of mortality 95% Confidence interval Variable (N = 2,250) Category General surgery Vascular surgery Surgical site infection Age, y Sex Female Male Ethnicity White Nonwhite ASA class 1 and 2 ASA class 3 and 4 Nonemergent operation Emergent operation

Odds ratio

Lower

Upper

P value*

Reference 1.52 1.06 1.03

0.64 0.34 1.01

3.61 3.24 1.05

.3420 .9230 .0100y

Reference 1.85

1.00

3.45

.0520z

0.19

3.94

.8490

1.61

31.35

.0100y

4.44

16.85

<.0001x

Reference 0.86 Reference 7.10 Reference 8.65

*Variables analyzed with a P value > .2 that are not shown this Table include the following: whether patients were admitted from their residence or another institution, diabetes, smoking status, steroid use, any wound class, and operative time. yDenotes statistical significance of P < .01. zDenotes statistical significance of P < .05. xDenotes statistical significance of P < .0001. ASA, American Society of Anesthesiologists.

After controlling for patient and intraoperative characteristics, developing an SSI was not a statistically significant factor in affecting a patient’s mortality (OR, 1.06; P = .923) (Table III). Hospital costs. Results of a linear regression analysis shown in Table IV suggest that the excess cost of an SSI was $10,497 (P = .003). In patients who had an SSI, male sex added $4,328 (P = .04) in costs, a nonwhite patient added $10,314 (P = .026), admission from residence added an additional $13,511 (P < .0001), steroid use added $9,064 (P = .019), and an ASA class 3 and 4 added $20,647 (P = .053). Also, an operative time between 4 and 8 hours added $10,798 (P < .0001), whereas an operative time greater than 8 hours added $10,180 (P = .004), and an emergent case added $17,498 (P < .0001). Other factors in the regression analysis did not affect the cost of SSIs. Effect on DOS. As seen in Table V, developing an SSI added 4.26 days (P < .0001) to the patient’s DOS. Many of the same factors that were significantly associated with costs also affected DOS. Male sex added 1.34 days (P = .03), admission from residence 7.02 days (P < .0001), steroid use 2.99 days (P = .006), ASA class 3 and 4 added 6.48 days (P = .032), contaminated and dirty wound classes 2.2 days (P = .009), operative time between 4 and 8 hours added 2.82 days (P < .0001), and emergent case 3.13 days (P < .0001).

DISCUSSION Given the continued occurrence of SSIs after general and vascular surgical procedures, it is important to understand their associated risk factors and costs to develop quality improvement programs tailored to prevention strategies and to make informed decisions regarding the costeffectiveness of these strategies. In this study of 2,250 surgical inpatients at a large academic medical center, independent risk factors for developing an SSI after general or vascular surgical procedures were male sex, insulin-dependent diabetes, wound classification, and operative time greater than 4 hours. Multiple studies in the literature have previously reported these clinical variables as risk factors, and this study adds support to the growing body of literature regarding SSIs. At our institution, patients with a cleancontaminated or dirty surgical procedure were 2 to 3 times more likely to develop a wound infection than clean procedures and 3 times more likely to have an SSI if the operation lasted greater than 8 hours. As in other studies, smoking and steroid use were not important risk factors in developing an SSI. In terms of mortality, although ASA class was not a factor in the development of an SSI, patients with an ASA class of 3 or 4 were greater than 7

Boltz et al 939

Surgery Volume 150, Number 5

Table IV. Linear regression of factors affecting cost of surgical site infection Variable (N = 2,250) Intercepty Surgical site infection Category General surgery Vascular surgery Sex Female Male Ethnicity White Nonwhite Patient transferred from another institution Patient admitted from residence No previous history of steroid use Steroid use ASA class 1 ASA class 2 ASA class 3 and 4 Operative time, h <4 4–8 >8 Nonemergent operation Emergent operation

Parameter estimate ($)

95% Confidence interval Lower

Upper

P value*

17,703 10,497

16,453 3,574

18,953 17,421

<.0001z .0030x

Reference 3,547

10,429

3,335

.3120

Reference 4,328

198

8,459

.0400{

1,232

19,396

.0260{

6,545

20,477

<.0001z

1,506

16,622

.0190{

8,862 272

32,895 41,567

.2590 .0530{

5,863 3,189

15,732 17,171

<.0001z .0040x

11,764

23,231

<.0001z

Reference 10,314 Reference 13,511 Reference 9,064 Reference 12,017 20,647 Reference 10,798 10,180 Reference 17,498

*Variables analyzed with a P value > .2 that are not shown in this Table include the following: age, diabetes, smoking status, and wound class. yIntercept represents hospital cost for patients without surgical site infections. zDenotes statistical significance of P < .0001. xDenotes statistical significance of P < .01. {Denotes statistical significance of P < .05. R2 = .26. This indicates that the model explains approximately 26% of overall variation in cost. ASA, American Society of Anesthesiologists.

times more likely to die after operation --- an outcome that reflects the serious nature of the associated comorbidities present at the time of operation in these patients. The presence of an SSI did not impact mortality unless it was an organspace infection. In addition, this study yielded a rate of SSI of 8.2%, which exceeds the current overall reported rate of approximately 4%.15 Most other studies rely on administrative databases, electronic markers of infection, or billing codes for SSI surveillance, which may underestimate the incidence of SSIs. In contrast, the NSQIP used in our study is unique, in that the database uses standard definitions of infection based on criteria established by the Centers for Disease Control and Prevention and clinical nurse reviewers to perform chart reviews to clinically detect SSIs, making the program more aggressive and probably more accurate at reporting SSIs retrospectively. This approach may account for

part of the difference between our SSI rate and the rate reported in the current literature. A study by Neumayer et al15 using data from the Department of Veterans Affairs’ NSQIP program reported an overall SSI rate of 4.3%. When the results of this study are broken down by procedure, however, general surgery patients who underwent procedures involving the stomach, intestine, appendix, biliary tract, or pancreas had an SSI rate of 40.6% and patients undergoing vascular procedures an SSI rate of 15.5%.15 Another study, this one carried out by Dimick et al,3 also used NSQIP data to assess postoperative occurrences of general and vascular surgery patients. This study reported an SSI rate of approximately 7%,3 further corroborating our rate of 8.2%. At our institution, the development of an SSI after a general or vascular surgical procedure incurs additional costs in excess of $10,497 for the hospital admission in which the procedure was

940 Boltz et al

Surgery November 2011

Table V. Linear regression of factors affecting duration of stay with surgical site infection

Variable (N = 2,250) Intercepty Surgical site infection Category General surgery Vascular surgery Sex Female Male Ethnicity White Nonwhite Patient transferred from another institution Patient admitted from their residence No previous history of steroid use Steroid use ASA class 1 ASA class 2 ASA class 3 and 4 Wound class Clean Clean-contaminated Contaminated and Dirty Operative time, h <4 4–8 >8 Nonemergent operation Emergent operation

Parameter estimate (d)

95% Confidence interval Lower

Upper

P value*

5.91 2.29

6.64 6.22

<.0001z <.0001z

Reference 0.57

2.52

1.38

.5660

Reference 1.34

0.16

2.51

.0260{

0.23

4.92

.0740

5.04

8.99

<.0001z

0.85

5.14

.0060z

1.20 0.55

10.65 12.42

.1180 .0320{

0.79 0.54

2.58 3.86

.2980 .0090x

1.42 0.77

4.22 3.20

<.0001z .2290

1.51

4.76

<.0001z

6.27 4.26

Reference 2.35 Reference 7.02 Reference 2.99 Reference 4.72 6.48 Reference 0.89 2.20 Reference 2.82 1.22 Reference 3.13

*Variables analyzed with a P value > .2 that are not shown in this Table include the following: age, diabetes, and smoking status. yIntercept represents duration of stay for patients without surgical site infections. zDenotes statistical significance of P < .0001. xDenotes statistical significance of P < .01. {Denotes statistical significance of P < .05. R2 = .26. This indicates that the model explains approximately 26% of overall variation in duration of stay. ASA, American Society of Anesthesiologists.

performed. Concomitantly, the DOS for a patient who develops an SSI is increased by 4.3 days. Importantly, our study highlights the financial implications in treating an SSI in select subpopulations of patients. Nonwhite patients who developed a wound infection incurred additional costs of $10,314, while it cost an additional $9,064 to treat an SSI in patients with a history of steroid use. Patients with an ASA class of 3 or 4 led to costs of $20,647 more than procedures for patients with an ASA class of 1, and operations exceeding 4 hours caused additional costs close to $10,000. These results demonstrate clearly the financial imperative in addition to the patient care imperative for decreasing the incidence of postoperative SSIs. These financial analyses assist in decisions whether to invest in potentially costly interventions

to prevent SSIs. For example, the 2% chlorhexidine/70% isopropyl alcohol preoperative scrub decreases SSIs by 41% compared to the povidoneiodine preoperative scrub; chlorhexidine-alcohol is also substantially more expensive than iodine products.16 Data on the costs of SSIs and the number of SSIs that may be prevented with chlorhexidine-alcohol can be used to assess the economic value of converting preoperative scrub products. Analyses must be undertaken with the understanding that no individual intervention will prevent all SSIs and that interventions implemented outside a research setting may not achieve results as dramatic as those obtained in a research setting. Lee et al17 estimated that if 2% chlorhexidine/ 70% isopropyl alcohol scrub led to a decrease in

Surgery Volume 150, Number 5

SSIs by at least 10%, net cost savings would be demonstrated. Our medical center is in the process of transitioning to chlorhexidine-alcohol as the standardized preoperative scrub primarily for efficacy reasons, but this decision appears to be supported by financial analyses. Although the economic impact of SSIs is acknowledged, few studies exist to quantify the costs to treat these infections or the financial determinants that contribute to those costs. In a literature review of the last 6 years, Broex et al18 could find only 16 articles that met their study criteria for reporting SSIs and economic evaluations and costs across all surgical disciplines. Studies specific to general and vascular surgical procedures are even rarer. Dimick et al3 used their NSQIP database to examine costs associated with postoperative complications at a single academic medical center. They demonstrated that their unadjusted total hospital costs for infectious complications were increased by $8,209, and DOS was increased by 4.0 days.3 When adjusted for complexity, patient variables, and other complications, their findings decreased to $1,398 of additional total hospital costs and 2.8 days for DOS; however, in their grouping of infectious complications, they included patients with sepsis and wound dehiscence. A study of SSIs in English hospitals by Coello et al19 examined the economic impact of wound infections in a number of different surgical procedures and specialties.19 They showed that superficial incisional SSIs were more common than deep incisional or organ-space SSIs, but the impact of organ-space SSIs on mortality was clinically important. In addition, all classes of SSIs contributed to significantly greater DOS and increased costs in every surgical procedure examined. The most in-depth attempt at examining the economic factors that contribute to the excess costs incurred by SSIs was published recently by Alfonso et al20 who found that 9% of their patients developed SSIs and that almost all the SSIs were concentrated in general, vascular, and cardiac patients. The overall DOS for these patients was increased by 14 days, and excess hospital costs were $10,232 per patient. Interestingly, in an attempt to include all costs, including both direct health costs and indirect social costs, incurred in treating SSIs, these investigators projected the overall costs of treating these infections from a societal perspective to be $97,433 in 2006 U.S. dollars. While all of these studies use slightly different study methods and surgical populations, they serve collectively to highlight the economic impact these

Boltz et al 941

SSIs have on the cost of delivering quality surgical care. It should be noted that the most important determinant of healthcare spending is the intensity of services provided over the course of a hospital stay. Therefore, the greatest driver of hospital costs is DOS. It can be assumed that the same variables or patient characteristics that contribute to greater DOS, such as sex or ethnicity (male and nonwhite, respectively, in our study), would also contribute to greater hospital costs. In addition, extended DOS relates to severity of illness. For example, a patient transferred from another institution is monitored constantly by healthcare staff and presents for an emergency procedure sooner than the patient admitted from their residence. Greater DOS is also affected by its surrogates of comorbid diseases and socioeconomic status.21 For example, Procter et al22 found that increased operative time is independently associated with a greater postoperative DOS after adjusting for type of procedure. One reason for this finding is that surgical procedures of greater duration have been shown clearly to induce major metabolic disorders that, in turn, cause postoperative morbidity. Another reason is that procedures of greater duration and DOS may be linked with factors affecting the patient’s underlying condition, such as nutritional status or other comorbid diseases.22 Although this study attempts to control for as many surrogate measures as possible within the confines of the dataset, future studies are needed to determine characteristics associated with duration of surgical procedures to identify modifiable variables and potentially decrease DOS and hospital costs. This study has several limitations. First, all of the patients studied were from a single academic medical center, and our results may not be representative of general and vascular surgery patients who develop SSIs in other institutions. Second, whereas this study addresses several patient factors affecting DOS, it does not account for socioeconomic status, which may prolong a hospital stay. In addition, the frequency of distribution of DOS can be skewed by a few patients whose stay is markedly prolonged. Nevertheless, DOS is a better reflection of resource utilization than the health status of the patient. Finally, the additional cost of $10,497 per patient reflects only hospital costs incurred as a result of the primary hospitalization. Total costs associated with SSI must include medical services, such as clinic visits, additional antibiotic use, home health services, readmission to the hospital, and

942 Boltz et al

additional procedures, as well as costs experienced directly by the patient, such as lost wages, transportation costs, and other out-of-pocket expenses. Therefore, our study almost certainly underestimates the real costs attributable to SSI for the surgical patient.

Surgery November 2011

11.

12.

REFERENCES 1. Malone DL, Genuit T, Tracy JK, Gannon C, Napolitano LM. Surgical site infections: reanalysis of risk factors. J Surg Res 2002;103:89-95. 2. Cardo D, Horan T, Andrus M, Dembinski M, Edwards J, Peavy G, et al. National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 2004;32:470-85. 3. Dimick JB, Chen SL, Taheri PA, Henderson WG, Khuri SF, Campbell DA Jr. Hospital costs associated with surgical complications: a report from the private-sector National Surgical Quality Improvement Program. J Am Coll Surg 2004; 199:531-7. 4. Khan NA, Quan H, Bugar JM, Lemaire JB, Brant R, Ghali WA. Association of postoperative complications with hospital costs and length of stay in a tertiary care center. J Gen Intern Med 2006;21:177-80. 5. Jenney AW, Harrington GA, Russo PL, Spelman DW. Cost of surgical site infections following coronary artery bypass surgery. ANZ J Surg 2001;71:662-4. 6. Pear R. Medicare says it won’t cover hospital errors. New York Times. August 19, 2007. http://www.nytimes.com/ 2007/08/19/washington/19hospital.html?sq=&st=nyt&ad xnnl=1&scp=1&adxnnlx=1304861137-/wm9DIA7esYD59ird BEObA. 7. Davenport DL, Henderson WG, Khuri SF, Mentzer RM Jr. Preoperative risk factors and surgical complexity are more predictive of costs than postoperative complications: a case study using the National Surgical Quality Improvement Program (NSQIP) database. Ann Surg 2005;242:463-8; discussion 468-71. 8. Fink AS, Campbell DA Jr, Mentzer RM Jr, Henderson WG, Daley J, Bannister J, et al. The National Surgical Quality Improvement Program in non-veterans administration hospitals: initial demonstration of feasibility. Ann Surg 2002;236:344-53; discussion 353-4. 9. Khuri SF. The NSQIP: a new frontier in surgery. Surgery 2005;138:837-43. 10. Khuri SF, Daley J, Henderson W, Hur K, Demakis J, Aust JB, et al. The Department of Veterans Affairs’ NSQIP: the first national, validated, outcome-based, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care. National VA

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

Surgical Quality Improvement Program. Ann Surg 1998; 228:491-507. Campbell DA Jr, Henderson WG, Englesbe MJ, Hall BL, O’Reilly M, Bratzler D, et al. Surgical site infection prevention: the importance of operative duration and blood transfusion–results of the first American College of Surgeons-National Surgical Quality Improvement Program Best Practices Initiative. J Am Coll Surg 2008;207:810-20. Shiloach M, Frencher S, Steeger J, Rowell K, Bartzokis K, Tomeh M. Toward Robust Information: Data Quality and Inter-Rater Reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2010;1:6-16. National Health Safety Network (HSN) modules. Patient safety component manual. Procedure-associated module. Surgical Site Infection (SSI) event: guidelines and procedures for monitoring SSI. October 2010;9:1–14. http://www.cdc. gov/nhsn/PDFs/pscManual/9pscSSIcurrent.pdf. Accessed January 7, 2010. Shwartz M, Young DW, Siegrist R. The ratio of costs to charges: how good a basis for estimating costs? Inquiry 1995;32:476-81. Neumayer L, Hosokawa P, Itani K, El-Tamer M, Henderson WG, Khuri SF. Multivariable predictors of postoperative surgical site infection after general and vascular surgery: results from the patient safety in surgery study. J Am Coll Surg 2007;204:1178-87. Darouiche RO, Wall MJ Jr, Itani KM, Otterson MF, Webb AL, Carrick MM, et al. Chlorhexidine-Alcohol versus Povidone-Iodine for Surgical-Site Antisepsis. N Engl J Med 2010;362:18-26. Lee I, Agarwal RK, Lee BY, Fishman NO, Umscheid CA. Systematic review and cost analysis comparing use of chlorhexidine with use of iodine for preoperative skin antisepsis to prevent surgical site infection. Infect Control Hosp Epidemiol 2010;31:1219-29. Broex EC, van Asselt AD, Bruggeman CA, van Tiel FH. Surgical site infections: how high are the costs? J Hosp Infect 2009;72:193-201. Coello R, Charlett A, Wilson J, Ward V, Pearson A, Borriello P. Adverse impact of surgical site infections in English hospitals. J Hosp Infect 2005;60:93-103. Alfonso JL, Pereperez SB, Canoves JM, Martinez MM, Martinez IM, Martin-Moreno JM. Are we really seeing the total costs of surgical site infections? A Spanish study. Wound Repair Regen 2007;15:474-81. Brasel KJ, Lim HJ, Nirula R, Weigelt JA. Length of stay: an appropriate quality measure? Arch Surg 2007;142:461-5; discussion 465-6. Procter LD, Davenport DL, Bernard AC, Zwischenberger JB. General surgical operative duration is associated with increased risk-adjusted infectious complication rates and length of hospital stay. J Am Coll Surg 2010;210:60-5.e1-2.